Explore the vast range of titles available.

The Lower Liaohe River Plain (LLRP) is a core grain production base in Northeast China. Monitoring the dynamics and changing rates of soil organic carbon (SOC) in cultivated lands is essential for regulating soil fertility, safeguarding food production, and maintaining the regional carbon balance. Based on soil survey data from three periods, 1980, 2008, and 2019, this study investigated the spatiotemporal dynamics of SOC content and its changing rate (SOCr) using geospatial analysis. Results showed that SOC content declined significantly from 11.19 g kg−1 to 10.47 g kg−1 during 1980–2008, then recovered slightly to 10.58 g kg−1 in 2019. Moreover, SOCr varied temporally in the period of 2008–2019, exhibiting a positive mean rate of 0.01 g kg−1 yr−1, which was significantly higher than that of the period of 1980–2008 (−0.03 g kg−1 yr−1). A significant negative correlation was examined between the initial SOC content and SOCr, showing an identification of the SOC equilibrium point (SOCep). The SOCep in the period of 2008–2019 was 9.69% higher than that in the period of 1980–2008. These findings provide a scientific basis for formulating regional policies and optimizing spatially differentiated management strategies to improve cropland SOC in the study area.

Maize straw returning is one of the important measures to improve dryland soil organic carbon (SOC). However, the effects of different maize parts on SOC fractions with different soil fertility levels in situ are not exactly clear. Therefore, an in situ field incubation experiment over 540 days, by adding different 13C-labeled maize parts (root, stem and leaf) into low- (no fertilizer treatment) and high- (manure treatment) fertility soils, was conducted at a long-term brown earth experimental site in Shenyang of China to figure out the effects of different maize parts on SOC fractions (dissolved organic carbon (DOC) and particulate organic carbon (POC)). The results showed that the distribution–DOC ratio of low-fertility treatment was higher than that of high-fertility treatment in the period of rapid decomposition of straw. In both low- and high-fertility soils, the ratio of carbon to DOC in leaf residue was higher than that in root and stem residues. The proportion of root, stem and leaf residue converted to DOC in low-fertility soil was 4.51%, 3.89% and 5.00%, respectively. The proportion of root, stem and leaf residue converted to DOC in high-fertility soil was 4.10%, 3.65% and 4.11%, respectively. As for the distribution–POC ratio, during the period of rapid decomposition of straw, the ratio of carbon conversion from root and stem residue to POC was generally higher than that from leaf residue. The ratio of carbon conversion to POC of root, stem and leaf residues in high-fertility treatment was higher than that in low-fertility treatment. In low-fertility treatment, the proportion of root, stem and leaf residues converted to POC was 41.34%, 46.33% and 36.11%, respectively. The proportion of root, stem and leaf residue converted to POC in high-fertility soil was 46.48%, 44.45% and 41.14%, respectively. The results showed that, for DOC, a low fertility level and more leaf residue types were beneficial. While, for POC, root and stem residues with a high fertility level were beneficial. These results provide evidence that the addition of different parts of maize residues would have differing effects on DOC and POC. Leaf residues in low-fertility soils were more suitable for increasing DOC. Root and stem residues in high-fertility soils were more suitable for increasing POC. Nevertheless, we could not ignore the unmeasured SOC fractions that some of the residues could be converted to.

The continuous data from the YL array and four AU stations during the lifetime of the cyclone TOMAS in March 2010 were downloaded from IRIS. By performing frequency-wave number (F-K) analysis on the array data, it was found that the orientation of the maximum energy of the secondary microseisms (0.1~0.5 Hz) was consistent with the movement of TOMAS when the central wind speed reached the typhoon level. The high wind speed of the cyclone could generate secondary microseisms as well as the high swell. However, the large global earthquake can affect the microseismic observations using F-K. The AU stations have a better microseism observation than the YL array, which might be due to the vibrating and tilting of the hydrophone caused by the turbulence. The F-K analysis on microseisms can produce better slowness and back azimuth observations than polarization.

Soil fungi play a crucial role in soil microbes, the composition and variety of whose communities can be altered due to nitrogen constraints, thereby affecting the plant’s development. This study aimed to investigate the relationship between the composition of soil fungi communities, fertility index, and the structure of soil fungal communities under varying nitrogen fertilizer conditions, using a long-term positioning test on the brown earth of Northeast China. It examined the impact of 31 years of applying of no fertilizer (CK, 0 kg N hm−2 a−1), the single application of inorganic fertilizer (N2, urea 135 kg N hm−2 a−1; N4, urea 270 kg N hm−2·a−1), the single application of organic fertilizer (M4, pig housing fertilizer 270 kg N hm−2 a−1), and mixed nitrogen fertilizer (M2N2, urea 135 N hm−2 a−1 + pig housing fertilizer 135 kg N hm−2 a−1) on the fertility index and fungal community structure of brown earth. The findings indicated the following: Long-term non-fertilization and the single application of chemical nitrogen fertilizer reduced the soil pH value and increased the soil bulk density. The application of organic fertilizer reduced soil bulk density and slowed down the reduction of soil fungal richness caused by nitrogen fertilizer application. The long-term application of different nitrogen fertilizers did not alter the dominant fungal phylum, showing that the dominant phylum in all treatments was Ascomycota. The pH, organic matter, total phosphorus, available phosphorus, total nitrogen, alkaline nitrogen, and available potassium were the main soil factors affecting the structural diversity of soil fungal communities. Total phosphorus explained the greatest differences in soil fungal communities.

Sugar transporters play important roles in controlling carbohydrate transport and are responsible for mediating the movement of sugars into cells in numerous organisms. In insects, sugar transporters not only play a role in sugar transport but may also act as receptors for virus entry and the accumulation of plant defense compounds. The brown planthopper, Nilaparvata lugens, inflicts damage on rice plants by feeding on their phloem sap, which is rich in sugars. In the present study, we identified 34 sugar transporters in N. lugens, which were classified into three subfamilies based on phylogenetic analysis. The motif numbers varied from seven to eleven, and motifs 2, 3, and 4 were identified in the functional domains of all 34 NlST proteins. Chromosome 1 was found to possess the highest number of NlST genes, harboring 15. The gut, salivary glands, fat body, and ovary were the different tissues enriched with NlST gene expression. The expression levels of NlST2, 3, 4, 7, 20, 27, 28, and 31 were higher in the gut than in the other tissues. When expressed in a Saccharomyces cerevisiae hexose transporter deletion mutant (strain EBY.VW4000), only ApST4 (previously characterized) and NlST4, 28, and 31 were found to transport glucose and fructose, resulting in functional rescue of the yeast mutant. These results provide valuable data for further studies on sugar transporters in N. lugens and lay a foundation for finding potential targets to control N. lugens.

Auditory Large Language Models (LLMs) have demonstrated strong performance across a wide range of speech and audio understanding tasks. Nevertheless, they often struggle when applied to low-resource or unfamiliar tasks. In case of labeled in-domain data is scarce or mismatched to the true test distribution, direct fine-tuning can be brittle. In-Context Learning (ICL) provides a training-free, inference-time solution by adapting auditory LLMs through conditioning on a few in-domain demonstrations. In this work, we first show that \\emph{Vanilla ICL}, improves zero-shot performance across diverse speech and audio tasks for selected models which suggest this ICL adaptation capability can be generalized to multimodal setting. Building on this, we propose \\textbf{Speech In-Context Learning Adaptation Training (SICL-AT)}, a post-training recipe utilizes only high resource speech data intending to strengthen model's in-context learning capability. The enhancement can generalize to audio understanding/reasoning task. Experiments indicate our proposed method consistently outperforms direct fine-tuning in low-resource scenario.

Dysarthria is a speech disorder characterized by impaired intelligibility and reduced communicative effectiveness. Automatic dysarthria assessment provides a scalable, cost-effective approach for supporting the diagnosis and treatment of neurological conditions such as Parkinson's disease, Alzheimer's disease, and stroke. This study investigates leveraging human perceptual annotations from speech synthesis assessment as reliable out-of-domain knowledge for dysarthric speech assessment. Experimental results suggest that such supervision can yield consistent and substantial performance improvements in self-supervised learning pre-trained models. These findings suggest that perceptual ratings aligned with human judgments from speech synthesis evaluations represent valuable resources for dysarthric speech modeling, enabling effective cross-domain knowledge transfer.

Conversational search systems require effective handling of context-dependent queries that often contain ambiguity, omission, and coreference. Conversational Query Reformulation (CQR) addresses this challenge by transforming these queries into self-contained forms suitable for off-the-shelf retrievers. However, existing CQR approaches suffer from two critical constraints: high dependency on costly external supervision from human annotations or large language models, and insufficient alignment between the rewriting model and downstream retrievers. We present ConvSearch-R1, the first self-driven framework that completely eliminates dependency on external rewrite supervision by leveraging reinforcement learning to optimize reformulation directly through retrieval signals. Our novel two-stage approach combines Self-Driven Policy Warm-Up to address the cold-start problem through retrieval-guided self-distillation, followed by Retrieval-Guided Reinforcement Learning with a specially designed rank-incentive reward shaping mechanism that addresses the sparsity issue in conventional retrieval metrics. Extensive experiments on TopiOCQA and QReCC datasets demonstrate that ConvSearch-R1 significantly outperforms previous state-of-the-art methods, achieving over 10% improvement on the challenging TopiOCQA dataset while using smaller 3B parameter models without any external supervision.

Large language models (LLMs) can adapt to new tasks through in-context learning (ICL) based on a few examples presented in dialogue history without any model parameter update. Despite such convenience, the performance of ICL heavily depends on the quality of the in-context examples presented, which makes the in-context example selection approach a critical choice. This paper proposes a novel Bayesian in-Context example Selection method (ByCS) for ICL. Extending the inference probability conditioned on in-context examples based on Bayes' theorem, ByCS focuses on the inverse inference conditioned on test input. Following the assumption that accurate inverse inference probability (likelihood) will result in accurate inference probability (posterior), in-context examples are selected based on their inverse inference results. Diverse and extensive cross-tasking and cross-modality experiments are performed with speech, text, and image examples. Experimental results show the efficacy and robustness of our ByCS method on various models, tasks and modalities.

This paper investigates the in-context learning abilities of the Whisper automatic speech recognition (ASR) models released by OpenAI. A novel speech-based in-context learning (SICL) approach is proposed for test-time adaptation, which can reduce the word error rates (WERs) with only a small number of labelled speech samples without gradient descent. Language-level adaptation experiments using Chinese dialects showed that when applying SICL to isolated word ASR, consistent and considerable relative WER reductions can be achieved using Whisper models of any size on two dialects, which is on average 32.3%. A k-nearest-neighbours-based in-context example selection technique can be applied to further improve the efficiency of SICL, which can increase the average relative WER reduction to 36.4%. The findings are verified using speaker adaptation or continuous speech recognition tasks, and both achieved considerable relative WER reductions. Detailed quantitative analyses are also provided to shed light on SICL's adaptability to phonological variances and dialect-specific lexical nuances.